Add flags to return scores, hidden states and / or attention weights in GenerationMixin (#9150)

* Define new output dataclasses for greedy generation

* Add output_[...] flags in greedy generation methods

Added output_attentions, output_hidden_states, output_scores flags in
generate and greedy_search methods in GenerationMixin.

* [WIP] Implement logic and tests for output flags in generation

* Update GreedySearchOutput classes & docstring

* Implement greedy search output accumulation logic

Update greedy_search unittests

Fix generate method return value docstring

Properly init flags with the default config

* Update configuration to add output_scores flag

* Fix test_generation_utils

Sort imports and fix isinstance tests for GreedySearchOutputs

* Fix typo in generation_utils

* Add return_dict_in_generate for backwards compatibility

* Add return_dict_in_generate flag in config

* Fix tyPo in configuration

* Fix handling of attentions and hidden_states flags

* Make style & quality

* first attempt attentions

* some corrections

* improve tests

* special models requires special test

* disable xlm test for now

* clean tests

* fix for tf

* isort

* Add output dataclasses for other generation methods

* Add logic to return dict in sample generation

* Complete test for sample generation

- Pass output_attentions and output_hidden_states flags to encoder in
encoder-decoder models
- Fix import satements order in test_generation_utils file

* Add logic to return dict in sample generation

- Refactor tests to avoid using self.assertTrue, which provides
scarce information when the test fails
- Add tests for the three beam_search methods: vanilla, sample and
grouped

* Style doc

* Fix copy-paste error in generation tests

* Rename logits to scores and refactor

* Refactor group_beam_search for consistency

* make style

* add sequences_scores

* fix all tests

* add docs

* fix beam search finalize test

* correct docstring

* clean some files

* Made suggested changes to the documentation

* Style doc ?

* Style doc using the Python util

* Update src/transformers/generation_utils.py

* fix empty lines

* fix all test

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

src/transformers/configuration_utils.py

@@ -124,6 +124,11 @@ class PretrainedConfig(object):
         - **num_return_sequences** (:obj:`int`, `optional`, defaults to 1) -- Number of independently computed returned
           sequences for each element in the batch that will be used by default in the :obj:`generate` method of the
           model.
+        - **output_scores** (:obj:`bool`, `optional`, defaults to :obj:`False`) -- Whether the model should return the
+          logits when used for generation
+        - **return_dict_in_generate** (:obj:`bool`, `optional`, defaults to :obj:`False`) -- Whether the model should
+          return a :class:`~transformers.file_utils.ModelOutput` instead of a :obj:`torch.LongTensor`
+
 
     Parameters for fine-tuning tasks
 
@@ -203,6 +208,8 @@ class PretrainedConfig(object):
         self.bad_words_ids = kwargs.pop("bad_words_ids", None)
         self.num_return_sequences = kwargs.pop("num_return_sequences", 1)
         self.chunk_size_feed_forward = kwargs.pop("chunk_size_feed_forward", 0)
+        self.output_scores = kwargs.pop("output_scores", False)
+        self.return_dict_in_generate = kwargs.pop("return_dict_in_generate", False)
 
         # Fine-tuning task arguments
         self.architectures = kwargs.pop("architectures", None)
@@ -343,6 +350,7 @@ class PretrainedConfig(object):
 
             Passing :obj:`use_auth_token=True` is required when you want to use a private model.
 
+
         Returns:
             :class:`PretrainedConfig`: The configuration object instantiated from this pretrained model.
 
@@ -372,6 +380,8 @@ class PretrainedConfig(object):
         From a ``pretrained_model_name_or_path``, resolve to a dictionary of parameters, to be used for instantiating a
         :class:`~transformers.PretrainedConfig` using ``from_dict``.
 
+
+
         Parameters:
             pretrained_model_name_or_path (:obj:`str` or :obj:`os.PathLike`):
                 The identifier of the pre-trained checkpoint from which we want the dictionary of parameters.